2426417
Description
Flashcards by tyson.schierholt, updated more than 1 year ago
|
Created by tyson.schierholt
over 9 years ago
|
|
| Question | Answer |
| What is multiple regression? | An extension of simple regression to the situation where there are 2 or more independent variables. |
| What is the general formula of a multiple regression line? | Y’ = a + b1X1 + b2X2 + ….. + bkXk |
| What indicates the unique effect of an independent variable, above and beyond the effect of other independent variables in the analysis? | Slope |
| A schematic diagram in which the variance of a variable is represented by the area of a circle, and the relationships among variables are represented by areas of overlap among circles. | Ballantine Diagram |
| In multiple regression, what needs to be done to calculate the slope of X1? | The effect of X2 needs to be removed. |
| What 2 equations are adjusted in multiple regression? | Slope (b) and Standard error (SEb). |
| As the relationship between Xs increases, does the standard error increase, stay the same, or decrease? | Increase |
| In multiple regression, what is the goodness of fit ratio called? | Squared multiple regression (R^2) |
| True/False: In general, the squared multiple correlation is equal to the sum of the two squared simple correlations. (R^2 = rY1^2 + rY2^2). | False (area b, the area shared between both squared simple correlations, is counted twice) |
| Which should be a larger value: squared multiple correlation (R^2) or adjusted squared multiple correlation (Radj^2)? | Squared multiple correlation (R^2) |
| How would you obtain a value of R^2 closer to the average R^2? | Repeat the analysis innumerable times using a series of similar samples. |
| Which provides a more realistic estimate of the goodness of fit: Squared multiple correlation or Adjusted squared multiple correlation? | Adjusted squared multiple correlation |
| Why is an Adjusted squared multiple correlation sometimes called a shrunken R^2? | It is always smaller than the sample R^2. |
| What is the degrees of freedom of standard error equal to? | Sample size (N) - # of X variables (k) - 1 |
| What is the situation called when the squared simple correlation is equal to 1.0? | Exact collinearity |
| What cannot be calculated if the squared simple correlation is equal to 1.0? | Standard error |
| What is the situation called when the simple squared correlation is near 1.0? | Multicollinearity |
| As the squared correlation between the independent variables increases, does the standard error of the slope increase, stay the same, or decrease? | Increase |
| As the standard error increases, does the confidence interval increase, stay the same, or decrease? | Increase |
| What condition makes it difficult to separate the effects of individual independent variables? | The independent variables become more and more correlated. |
| True/False: There is always some uncertainty about the true slope of each independent variable. | True |
| What are the 2 ways to address the uncertainty around the slope? | 1. Calculate the t ratio for each slope 2. Calculate the confidence intervals for each slope |
| What is the way to address the uncertainty of the squared multiple correlation (R^2)? | Calculate the F ratio |
| What represents the overall effect of the independent variables? | Goodness of fit ratio (R^2) |
| If SE1 has a SS1 4x greater than SS2, what is the value of SE2 relative to SE1? | SE2 = 2 x SE1 |
| If you have a 90% confidence interval, what is the percentage for alpha? | 10% |
| As the value of Fcalc decreases, does the value of tcalc increase, stay the same, or decrease? | Decrease |
| When do you use the adjusted squared multiple correlation (Radj^2) instead of the squared multiple correlation (R^2)? | Radj^2 is much smaller than R^2. |
| What is df1 and df2 for Fcalc? | df1 = k df2 = N - k - 1 |
| Which do you use if there is no clear evidence of a variable's effect: Simple regression line or Multiple regression line? | Simple regression line (Since only 1 variable has a statistically significant effect on the dependent variable) |
| What does a curvilinear regression line allow the goodness of fit to do? | It allows the goodness of fit to fit the sample better in some cases. |
| What is the equation for a curvilinear regression line? | Y' = a + b1X + b2X^2 |
| Y' = a + b1X + b2X^2 Which term in this equation allows the curved line of best fit? | b2X^2 |
| How many independent variables are in the curvilinear regression line? | 1 independent variable |
| As SP12 and r12 increase, does tcalc increase, stay the same, or decrease? | Decrease |
| The likelihood of a correct positive. | Power |
| True/False: Standardization of slope gives a mean of 0 and a standard deviation of 1. | True |
| A standardized slope expresses the effect of an independent variable in terms of what? | In terms of z scores. |
| Why is the constant omitted from the equation for the standardized regression line? | The constant is always 0 for a standardized regression line. |
| What is the value for standardized sums of squares? | SS = N - 1 |
| What are the values of the means for each of the variables for a standardized regression line? | Always 0 since it is a mean of z scores. |
| How do you standardize the slope? | Bk = bk x sk/sY Bk = bk x sqrt(SSk/SSY) |
| True/False: The equations for a non-standardized regression line can be used with the standardized SS and SP. | True |
| True/False: The standardized tcalc is equal to the non-standardized tcalc. | True |
| True/False: The issue of statistical significance depends on the units. | False |
| How do you find the constant in multiple regression with categorical variables? | a = M0 |
| How do you find the slope of X2 in multiple regression with categorical variables? | b2 = M2 - M0 |
| How many columns of dummy coding do you need for 4 categorical variables? | 3 (Always 1 less than the number of X variables) |
Want to create your own Flashcards for free with GoConqr? Learn more.